CN106503094A - A kind of user preference analysis method based on document - Google Patents
A kind of user preference analysis method based on document Download PDFInfo
- Publication number
- CN106503094A CN106503094A CN201610896081.7A CN201610896081A CN106503094A CN 106503094 A CN106503094 A CN 106503094A CN 201610896081 A CN201610896081 A CN 201610896081A CN 106503094 A CN106503094 A CN 106503094A
- Authority
- CN
- China
- Prior art keywords
- document
- user
- value
- user preference
- style
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/93—Document management systems
Abstract
The invention discloses a kind of user preference analysis method based on document, including:Receive user document;The corresponding informance of each dimension in customer documentation according to default document analysis model extraction is simultaneously calculated, so as to obtain the total preference value of user;Wherein, the document analysis model is:H (x)=V1S1+V2S2+…+VnSn, wherein, H (x) represents user for the total preference value of the user of particular aspects, V1、V2…VnThe user preference value of each dimension of the particular aspects is represented respectively;S1、S2…SnRepresent the corresponding default weight coefficient of each dimension, n >=2;The total preference value of the user for having preserved is updated using the total preference value of the calculated user.A kind of user preference analysis method based on document disclosed by the invention has filled up Behavior preference analysis in the blank in document field, provides new source for personalized ventilation system resource.
Description
Technical field
A kind of the present invention relates to user behavior preference analysis field, more particularly to user preference analysis side based on document
Method.
Background technology
With the gradually popularization of the Internet and becoming increasingly abundant for network information resource, people are gradually walked from the absence of information epoch
Enter the information overload epoch, it is very big that the continuous increase of quantity of information causes either information producer or information consumer all to encounter
Difficult and challenge:The information for finding oneself demand from the information of magnanimity has become more and more difficult.At the same time, want to allow
Product is shown one's talent in numerous information, and the concern for obtaining consumer is also more and more difficult.The method that commonly used in the past is to allow user
Oneself input keyword search, search engine can be traveled through in lane database according to the key word submitted to by user, suitable to find
Information recommended.The shortcoming of this method is:The clear and definite demand of oneself of user is needed, and needs user actively to retrieve.
In order to high efficiency, quick, actively provide the user with recommendation information, excavate user's information needed from mass data
And recommend to user, the time that user obtains useful information is saved, personalized ventilation system arises at the historic moment, personalized ventilation system
Corresponding user behavior preference analysis can be carried out according to user behavior data first, conventional method is by user behavior analysis
Set up a user preferences modeling, the preference that the behavior of user is converted to user.Mostly modeling to user behavior preference at present it is
In information search engine, net purchase platform etc. to user information search keyword, search information, net purchase merchandise news, net purchase evaluation
The aspects such as information carry out preference analysis.
The present inventor's discovery when the research and development present invention, is also not based on the user behavior preference of document at present on the market
Analysis method and user behavior preference analysis model, but user has spent in plenty of time and the elaborate document of energy and has included
The bulk information of user behavior preference, selection of the such as user to document font, the setting of document color, document characters in a fancy style
Use etc., the Behavior preference of user can be reflected, which results in the Internet document user behavior preference analysis this
The blank and the Internet document resources in field are using a kind of upper waste.
Content of the invention
The purpose of the embodiment of the present invention is to provide a kind of user preference analysis method based on document, can enrich user's row
The utilization rate of source database and raising the Internet document resources for preference analysis, while fill up user behavior preference analysis exist
The blank in this field of document.
For achieving the above object, a kind of user preference analysis method based on document is embodiments provided, including:
Receive user document;
The corresponding informance of each dimension in customer documentation according to default document analysis model extraction is simultaneously counted
Calculate, so as to obtain the total preference value of user;Wherein, the document analysis model is:H (x)=V1S1+V2S2+V3S3+V4S4+…+
VnSn, wherein, H (x) represents user for the total preference value of the user of particular aspects, V1、V2、……VnThe certain party is represented respectively
The user preference value of each dimension in face, the user preference value of each dimension is based on the corresponding informance for obtaining and by default
Formula is calculated;S1、S2、……SnRepresent the corresponding default weight coefficient of each dimension, n >=2;
The total preference value of the user for having preserved is updated using the total preference value of the calculated user.
Compared with prior art, customer documentation is divided into multiple by the embodiment of the present invention according to default document analysis model
Dimension, extracts the document corresponding informance of each dimension, and the document corresponding informance based on default formula to each dimension
The user preference value of each dimension is calculated, so that user is obtained for the user of particular aspects (such as file design style)
Total preference value is simultaneously updated.Newest user is obtained based on the user preference analysis method of document by what the present embodiment was provided
After the total preference value of user of particular aspects (such as file design style), can be carried using the total preference value of newest user
For corresponding personalized service resource (for example, there is provided meet Service Source of the user in the preference of file design style).Therefore,
The source database that can enrich user behavior preference analysis based on the user preference analysis method of document that the present embodiment is provided,
The utilization rate of the Internet document resources is improved, blank of the user behavior preference analysis in this field of document has also been filled up, while
New source is provided for personalized ventilation system resource.
As the improvement of such scheme, also include step:
In Push Service resource, according to renewal after the total preference value of the user push and meet user in the certain party
The Service Source of the preference in face.
Used as the improvement of such scheme, the particular aspects include that file design style, the file design style include
Document font, document color, three dimensions of document characters in a fancy style, the corresponding informance of each dimension in the customer documentation include text
Shelves font information, document color information and document characters in a fancy style information.
As the improvement of such scheme, pre-set the total preference value of user of the file design style max-thresholds and
Minimum threshold, the corresponding file design style of the max-thresholds are that pole complicates style, the corresponding document of the minimum threshold
For extremely simplifying style, each value between the max-thresholds and the minimum threshold corresponds to the extremely complicated wind-transformation to design style
Each file design style between lattice and extremely simplified style.
As the improvement of such scheme, the document font information is extracted by following steps and the corresponding user of calculating is inclined
Good value:
The font of each character in document is obtained, the character number of corresponding font used in statistic document, by formula P=
(a1*b1+a2*b2+…+ai*bi)/(b1+b2+…bi) calculate document in character average thickness value, wherein, a1、a2…aiRepresent
The thickness value of specific font;b1、b2…biRepresent the character number using corresponding font, P represents the average of character in the document
Thickness value, i >=1;
The user preference value that document font is calculated by formula P (f)=PQ, wherein:P (f) represents the document font
User preference value, Q represent thickness weight coefficient, Q=| 1.5-P |;
User preference value P (f) is normalized to the threshold range between the max-thresholds and the minimum threshold
Interior, so as to obtain user preference value V1.
As the improvement of such scheme, the document color information is extracted by following steps and the corresponding user of calculating is inclined
Good value:
To the document, per page carries out sectional drawing, obtains the document each pixel by carrying out image procossing to every width sectional drawing
Depth angle value, so as to obtain the depth angle value of each color in the document, and pass through formula P (c)=Y1A1+Y2A2+Y3A3+
Y4A4+…+YjAjThe user preference value of document color is calculated, wherein:P (c) represents the user preference value of the document color, Y1、
Y2……YjRepresent the depth angle value of each color, A1、A2……AjRepresent that each shade angle value accounts for the area of whole document
Ratio, j >=1;
User preference value P (c) is normalized to the threshold range between the max-thresholds and the minimum threshold
Interior, so as to obtain user preference value V2.
As the improvement of such scheme, the area ratio that each shade angle value accounts for whole document is calculated by following steps
Example:
To the document, per page carries out sectional drawing, obtains the document each pixel by carrying out image procossing to every width sectional drawing
Depth angle value;
By the ratio of cumulative for the depth angle value identical pixel rear and total pixel of document, account for as each shade angle value whole
The area ratio of individual document.
As the improvement of such scheme, by formula Yj=Rm+Gb+Bk calculates the depth angle value of each color, wherein:Yj
Represent that the depth angle value of each color in the document, R, G, B represent the numerical value of RGB channel in each pixel, m, b, k respectively
Represent the coefficient of the RGB channel.
As the improvement of such scheme, m=0.299, b=0.587, k=0.114.
As the improvement of such scheme, the document characters in a fancy style information is extracted by following steps and calculate corresponding user
Preference value:
The character number of each characters in a fancy style in document is obtained, by formula P (a)=U1C1+U2C2+U3C3…+UtCtCalculate
The user preference value of document characters in a fancy style, wherein:P (a) represents the user preference value of the document characters in a fancy style, U1、U2……UtRepresent
The character number of each characters in a fancy style, C1、C2……CtRepresent the default preference weight system of each characters in a fancy style corresponding
Number, t >=1;
User preference value P (a) is normalized to the threshold range between the max-thresholds and the minimum threshold
Interior, so as to obtain user preference value V3.
Description of the drawings
Fig. 1 is a kind of flow chart of the user preference analysis method based on document in the embodiment of the present invention 1;
Fig. 2 is the idiographic flow schematic diagram of step S2 in Fig. 1;
Fig. 3 is that in step S201, the total preference value threshold value of document design style user arranges schematic diagram in Fig. 2;
Fig. 4 is the idiographic flow schematic diagram of step S202 in Fig. 2;
Fig. 5 is the idiographic flow schematic diagram of step S203 in Fig. 2;
Fig. 6 is the idiographic flow schematic diagram of step S204 in Fig. 2;
Fig. 7 is a kind of flow chart of the user preference analysis method based on document in the embodiment of the present invention 2.
Specific embodiment
Accompanying drawing in below in conjunction with the embodiment of the present invention, to the embodiment of the present invention in technical scheme carry out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiment.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
Referring to Fig. 1, it is a kind of flow chart of user preference analysis method based on document that the embodiment of the present invention 1 is provided.
A kind of user preference analysis method based on document that the embodiment of the present invention 1 is provided includes step S1~S3:
S1:Receive user document.
When being embodied as, reception can be user's real-time manufacturing and synchronously preserve to the document in the Internet high in the clouds, or
It is user's real-time manufacturing for carrying out the document of user preference analysis, can also be that user has been stored in local or high in the clouds
And extract for carrying out the document of user preference analysis.In the present embodiment, in order in time and to automatically obtain user newest
Preference, preferably receive user real-time manufacturing the document that synchronously preserves to the Internet high in the clouds is for carrying out user preference point
Analysis.
S2:The corresponding informance of each dimension in customer documentation according to default document analysis model extraction is simultaneously carried out
Calculate, so as to obtain the total preference value of user;Wherein, the document analysis model is:H (x)=V1S1+V2S2+V3S3+V4S4+…+
VnSn, wherein, H (x) represents user for the total preference value of the user of particular aspects, V1、V2、……VnThe certain party is represented respectively
The user preference value of each dimension in face, the user preference value of each dimension is based on the corresponding informance for obtaining and by default
Formula is calculated;S1、S2、……SnRepresent the corresponding default weight coefficient of each dimension, n >=2.
User behavior preference analysis model and computational methods based on document are provided in this step, specifically, at this
In step, total preference value H (x) of the user of the particular aspects includes but is not limited to file design style, the file design wind
The dimension of lattice includes but is not limited to document font, document color, document characters in a fancy style, and the corresponding informance of document font dimension includes text
Shelves font information, the corresponding informance of document color dimension include that document color information, the corresponding informance of document characters in a fancy style include text
Shelves characters in a fancy style information.
S3:The total preference value of the user for having preserved is updated using the total preference value of the calculated user.
Specifically, after document every time using document analysis model analysiss user, will every time calculated user total
Preference value is updated, and the total preference value result of calculation of user that will be newest is updated to the total preference value of the user for having preserved.
Customer documentation is divided into multiple dimensions according to default document analysis model by the embodiment of the present invention 1, is extracted each
The document corresponding informance of individual dimension, and each dimension is calculated based on default formula to the document corresponding informance of each dimension
The user preference value of degree, so as to obtaining user for the total preference value of the user of particular aspects (such as file design style) and carrying out
Update.Newest user is obtained in particular aspects (example based on the user preference analysis method of document by what the present embodiment 1 was provided
Such as file design style) the total preference value of user after, corresponding individual character can be provided using the total preference value of newest user
Change Service Source (for example, there is provided meet Service Source of the user in the preference of file design style).Therefore, the embodiment of the present invention
1 source database that can be enriched user behavior preference analysis based on the user preference analysis method of document for providing, is improved mutually
The utilization rate of networking document resources, has also filled up blank of the user behavior preference analysis in this field of document, while being individual character
Change recommendation service resource and provide new source.
Understand the corresponding informance of each dimension in the foundation of document analysis model and customer documentation in step S2 for convenience
Extraction process, name file design style, as a example by dimension is respectively document font, document color, document characters in a fancy style, in detail
Illustrate how to calculate the total preference value of user for determining user in file design style.
Referring to Fig. 2, by calculate determine user file design style the total preference value of user include step S201~
S205, wherein:
S201:The max-thresholds and minimum threshold of the total preference value of user of the file design style are pre-set, described
The corresponding file design style of max-thresholds is that pole complicates style, and the corresponding file design style of the minimum threshold is extremely letter
Wind-transformation lattice, the corresponding pole of each value between the max-thresholds and the minimum threshold complicate style and extremely simplify style
Between each file design style.
Specifically, the brief or complicated of the file design style is intuitively represented in order to convenient, referring to Fig. 3, Fig. 3
Be in Fig. 2 in step S201 document design style threshold value arrange schematic diagram, by total for user preference value arrange minimum threshold be 0, most
Big threshold value is 100, and minimum threshold 0 is corresponding extremely to simplify style, and the corresponding pole of max-thresholds 100 complicates style, and minimum threshold 0 is to most
Each value between big 100 interval of threshold value corresponds to described each file design extremely simplified between style and pole complication style
Style.It should be understood that the concrete numerical value that the minimum threshold and max-thresholds are arranged can be set according to practical situation and demand
Put, such as may also set up minimum threshold for 0, max-thresholds is 10, and here is not particularly limited.
S202:The document font information in the document is extracted, user is calculated for document font according to preset formula
User preference value, and calculated user preference value is normalized to the threshold between the max-thresholds and the minimum threshold
In the range of value.
Specifically, referring to figure, 4, the step can be realized by following steps, including step S2021~S2023, wherein:
S2021:The font of each character in document is obtained, the character number of corresponding font, passes through used in statistic document
Formula P=(a1*b1+a2*b2+…+ai*bi)/(b1+b2+…bi) calculate document in character average thickness value, wherein, a1、a2…
aiRepresent the thickness value of specific font;b1、b2…biRepresent that the character number using corresponding font, P represent character in the document
Average thickness value, i >=1.
Specifically, each font all corresponds to a thickness value, and this is written in every kind of font, can be with direct access
Arrive, the thickness value that can obtain each character in document by the font for obtaining each character in document has in statistic document
The character number of identical thickness value, can be calculated the average thickness value of character in document.
S2022:The user preference value that document font is calculated by formula P (f)=PQ, wherein:P (f) represents the document
The user preference value of font, Q represent thickness weight coefficient, Q=| 1.5-P |.
Specifically, what thickness weight coefficient Q was represented is its impact feelings to document style for some average thickness value
Condition, each thickness value correspond to a thickness weight coefficient, and the font weight value deflection of concision style is less, and definition is average thick
When thin value is 1.5, the requirement of concision style is best suited.
S2023:User preference value P (f) is normalized to the threshold between the max-thresholds and the minimum threshold
In the range of value, so as to obtain user preference value V1.
The step is set for calculated document font user preference value P (f) to be normalized to the document for pre-setting
In the total max-thresholds of preference value of user and the threshold range of minimum threshold of meter style so that document font this dimension user
Preference value result of calculation is in the range of the total preference value of file design style user so that document font user preference value is set with document
The total preference value of meter style user has unified criterion.
Specifically, as each font all corresponds to a thickness value (thickness value scope is fixed), such as with 1.5-2.0 point
The thickness value scope of each font is not represented, 1.5 is minimum, 2.0 is maximum.Can be obtained according to the computing formula in step S2021
To in the document, the scope of average thickness value P of character is also 1.5~2.0, therefore obtains document by formula P (f)=PQ
The scope of user preference value P (f) of font is 0~1.Then, user preference value P (f) that scope is 0~1 is returned by mapping
One max-thresholds for changing the total preference value of user to the file design style for pre-setting and the threshold range (example of minimum threshold
Such as, 0~100) in, normalize to user preference value V in the threshold range so as to obtain user preference value P (f)1.
S203:The document color information in the document is extracted, user is calculated for document font according to preset formula
User preference value, and calculated user preference value is normalized to the threshold between the max-thresholds and the minimum threshold
In the range of value.
Specifically, referring to Fig. 5, the step can be realized by following steps, including step S2031~S2032, wherein:
S2031:To the document, per page carries out sectional drawing, obtains the document often by carrying out image procossing to each sectional drawing
The depth angle value of individual pixel, so as to obtain the depth angle value of each color in the document, and passes through formula P (c)=Y1A1+
Y2A2+Y3A3+Y4A4+…+YjAjThe user preference value of document color is calculated, wherein:P (c) represents that the user of the document color is inclined
Good value, Y1、Y2……YjRepresent the depth angle value of each color, A1、A2……AjRepresent that each shade angle value accounts for whole text
The area ratio of shelves, j >=1.
Specifically, the document integral color of concision style is partially light, and to document, per page carries out sectional drawing, to document per cut
Figure carries out image procossing, obtains the depth angle value of each pixel, so as to obtain the depth angle value of each color in the document, tool
Body, by formula Yj=Rm+Gb+Bk calculates the depth angle value of each color, wherein:YjRepresent each face in the document
The depth angle value of color, R, G, B represent that the numerical value of RGB channel in each pixel, m, b, k represent the coefficient of the RGB channel respectively,
M=0.299, b=0.587, k=0.114.The numerical value of RGB channel is different, then the depth angle value of pixel is different, i.e. each color
Depth angle value different, a kind of each depth angle value corresponding color respectively then can be obtained by the depth angle value of each pixel
To all colours that document is used, and the corresponding depth angle value of each color, by will be identical for each depth angle value described
Pixel carry out cumulative obtain accumulated value, accumulated value is each shade angle value with the ratio of the total pixel of document and accounts for whole text
The area ratio of shelves.
S2032:User preference value P (c) is normalized to the threshold between the max-thresholds and the minimum threshold
In the range of value, so as to obtain user preference value V2.
The step is set for calculated document color user preference value P (c) to be normalized to the document for pre-setting
In the total max-thresholds of preference value of user and the threshold range of minimum threshold of meter style so that document color this dimension user
Preference value result of calculation is in the range of the total preference value of file design style user so that document color user preference value is set with document
The total preference value of meter style user has unified criterion.
Specifically, the numerical value of the RGB channel of black for (0,0,0), white RGB channel numerical value for (255,255,
255), according to formula Y in step S2031jIt is 0~255 that=Rm+Gb+Bk can obtain the scope of the depth angle value of each color,
According to formula P (c) in step S2031=Y1A1+Y2A2+Y3A3+Y4A4+…+YjAjDocument color user preference value P can be obtained
C the scope of () is 0~255, user preference value P (c) that scope is 0~255 is normalized to the text for pre-setting by mapping
The max-thresholds and minimum threshold of the total preference value of user of shelves design style threshold range (for example, 0~100) in, so as to
User preference value V in the threshold range is normalized to user preference value P (c)2.
S204:The document characters in a fancy style information in the document is extracted, user is calculated for document font according to preset formula
User preference value, and calculated user preference value is normalized between the max-thresholds and the minimum threshold
In threshold range
Specifically, referring to Fig. 6, the step can be realized by following steps, including step S2041~S2042, wherein:
S2041:The character number of each characters in a fancy style in document is obtained, by formula P (a)=U1C1+U2C2+U3C3…+
UtCtThe user preference value of document characters in a fancy style is calculated, wherein:P (a) represents the user preference value of the document characters in a fancy style, U1、
U2……UtThe character number of each characters in a fancy style, C described in representing1、C2……CtRepresent the default of each characters in a fancy style corresponding
Preference weight coefficient, t >=1;
Specifically, the species of characters in a fancy style used in document and the character using each characters in a fancy style corresponding are determined
Number, is to determine due to the effect of characters in a fancy style, and each characters in a fancy style has corresponding style attribute, rule of thumb defined each
Plant the default preference weight coefficient of characters in a fancy style.
S2042:User preference value P (a) is normalized to the threshold between the max-thresholds and the minimum threshold
In the range of value, so as to obtain user preference value V3.
The step is used for calculated document characters in a fancy style user preference value P (a) to be normalized to the document for pre-setting
In the total max-thresholds of preference value of the user of design style and the threshold range of minimum threshold so that document characters in a fancy style this dimension
User preference value result of calculation is in the range of the total preference value of file design style user so that document characters in a fancy style user preference value with
The total preference value of file design style user has unified criterion.
Specifically, due to analyze document in character number be to determine, limited, then used in document characters in a fancy style species
And the use of the character number of each characters in a fancy style corresponding is also limited, according to the calculated text of formula in step S2041
The user preference value scope of shelves characters in a fancy style is also limited, sets user preference value P (a) scope of document characters in a fancy style as 0~1, and 0
Represent, 1 represents that character is all using the characters in a fancy style of pole complication style in document, then, by model
Enclosing, the total preference value of user of the file design style for pre-setting is normalized to by mapping for 0~1 user preference value P (a)
Max-thresholds and minimum threshold threshold range (for example, 0-100) in, so as to obtain user preference value P (a) normalization
User preference value V Dao the threshold range in3.
S205:User preference value V that step S202~S204 is respectively obtained1, user preference value V2, user preference value V3
Substitute into document analysis model H (x)=V1S1+V2S2+V3S3In, so as to be calculated user for the user of file design style is total
Preference value.
In this step, H (x) represents user for the total preference value of the user of file design style, V1、V2、V3Represent respectively
User for the document font in file design style, document color, three dimensions of document characters in a fancy style user preference value, and
S1、S2、S3Represent the corresponding default weight coefficient of each dimension.It should be understood that due to S1、S2、S3Represent that each dimension is corresponding
Default weight coefficient, S1、S2、S3Three coefficients and be 1.
Referring to Fig. 7, it is a kind of flow chart of user preference analysis method based on document that the embodiment of the present invention 2 is provided.
A kind of user preference analysis method based on document that the embodiment of the present invention 2 is provided includes step S21~S24:
S21:Receive user document.
When being embodied as, reception can be user's real-time manufacturing and synchronously preserve to the document in the Internet high in the clouds, or
It is user's real-time manufacturing for carrying out the document of user preference analysis, can also be that user has been stored in local or high in the clouds
And extract for carrying out the document of user preference analysis.In the present embodiment, in order in time and to automatically obtain user newest
Preference, preferably receive user real-time manufacturing the document that synchronously preserves to the Internet high in the clouds is for carrying out user preference point
Analysis.
S22:The corresponding informance of each dimension in customer documentation according to default document analysis model extraction is gone forward side by side
Row is calculated, so as to obtain the total preference value of user;Wherein, the document analysis model is:H (x)=V1S1+V2S2+V3S3+V4S4+…
+VnSn, wherein, H (x) represents user for the total preference value of the user of particular aspects, V1、V2、……VnRepresent described specific respectively
The user preference value of each dimension of aspect, the user preference value of each dimension is based on the corresponding informance for obtaining and by pre-
If formula is calculated;S1、S2、……SnRepresent the corresponding default weight coefficient of each dimension, n >=2.;
Specifically, in this step, total preference value H (x) of the user of the particular aspects includes but is not limited to file design
Style, the dimension of the file design style include but is not limited to document font, document color, document characters in a fancy style, document font
The corresponding informance of dimension includes that document font information, the corresponding informance of document color dimension include document color information, document skill
The corresponding informance of art word includes document characters in a fancy style information.
S23:The total preference value of the user for having preserved is updated using the total preference value of the calculated user.
Specifically, after document every time using document analysis model analysiss user, will every time calculated user total
Preference value is updated, and the total preference value result of calculation of user that will be newest is updated to the total preference value of the user for having preserved.
S24:In Push Service resource, according to renewal after the total preference value of the user push and meet user in the spy
The Service Source of the preference of fixed aspect.
Specifically, the step is used for for total for document analysis user preference value result of calculation being applied to personalized ventilation system neck
In domain, the total preference value of the calculated user of the according to embodiments of the present invention 2 document analysis models for providing can be pushed to user
Service Source, for example, provide resource supplying at aspects such as shopping, viewings to client.
The embodiment of the present invention 2 provide a kind of user preference analysis method based on document in step S21~S23 with this
In a kind of user preference analysis method based on document that bright embodiment 1 is provided, step S1~S3 is identical, in the embodiment of the present invention 2
Step S21~S23 specific implementation process is identical with step S1~S3 in the embodiment of the present invention 1, will not be described here.
Compared with Example 1, a kind of user preference analysis method based on document that the embodiment of the present invention 2 is provided increased
Step S24, when being embodied as, the step is used for for total for document analysis user preference value result of calculation being applied to personalized recommendation clothes
In business field, the total preference value of the calculated user of the according to embodiments of the present invention 2 document analysis models for providing can be to user
Push Service resource, for example, provide resource supplying at aspects such as shopping, viewings to client.
Compared with Example 1, a kind of user preference analysis method based on document that the embodiment of the present invention 2 is provided is by basis
The total preference value result of calculation of the calculated user of document analysis model analysiss that the present invention sets up is applied to personalized recommendation clothes
The source database of personalized recommendation in business field, is enriched, and new source is provided for personalized ventilation system resource.
The above is the preferred embodiment of the present invention, it is noted that for those skilled in the art
For, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications are also considered as
Protection scope of the present invention.
Claims (10)
1. a kind of user preference analysis method based on document, it is characterised in that include:
Receive user document;
The corresponding informance of each dimension in customer documentation according to default document analysis model extraction is simultaneously calculated, from
And obtain the total preference value of user;Wherein, the document analysis model is:H (x)=V1S1+V2S2+V3S3+V4S4+…+VnSn, its
In, H (x) represents user for the total preference value of the user of particular aspects, V1、V2、……VnThe particular aspects every is represented respectively
The user preference value of individual dimension, the user preference value of each dimension based on the corresponding informance for obtaining and pass through preset formula meter
Obtain;S1、S2、……SnRepresent the corresponding default weight coefficient of each dimension, n >=2;
The total preference value of the user for having preserved is updated using the total preference value of the calculated user.
2. a kind of user preference analysis method based on document as claimed in claim 1, it is characterised in that also include step:
In Push Service resource, according to renewal after the total preference value of the user push and meet user in the particular aspects
The Service Source of preference.
3. a kind of user preference analysis method based on document as claimed in claim 1, it is characterised in that the particular aspects
Including file design style, the file design style includes document font, document color, three dimensions of document characters in a fancy style, institute
The corresponding informance for stating each dimension in customer documentation includes document font information, document color information and document characters in a fancy style letter
Breath.
4. a kind of user preference analysis method based on document as claimed in claim 3, it is characterised in that pre-set described
The max-thresholds and minimum threshold of the total preference value of the user of file design style, the corresponding file design style of the max-thresholds
Complicate style for pole, the corresponding file design style of the minimum threshold for extremely simplifying style, max-thresholds and described
Each file design style between the corresponding pole complication style of each value and extremely simplified style between minimum threshold.
5. a kind of user preference analysis method based on document as claimed in claim 4, it is characterised in that by following steps
Extract the document font information and calculate corresponding user preference value:
The font of each character in document is obtained, the character number of corresponding font used in statistic document, by formula P=(a1*
b1+a2*b2+…+ai*bi)/(b1+b2+…bi) calculate document in character average thickness value, wherein, a1、a2…aiRepresent specific
The thickness value of font;b1、b2…biRepresent that the character number using corresponding font, P represent the average thickness of character in the document
Value, i >=1;
The user preference value that document font is calculated by formula P (f)=PQ, wherein:P (f) represents the user of the document font
Preference value, Q represent thickness weight coefficient, Q=| 1.5-P |;
User preference value P (f) is normalized in the threshold range between the max-thresholds and the minimum threshold, from
And obtain user preference value V1.
6. a kind of user preference analysis method based on document as claimed in claim 3, it is characterised in that by following steps
Extract the document color information and calculate corresponding user preference value:
To the document, per page carries out sectional drawing, by carrying out the depth that image procossing obtains each pixel of document to every width sectional drawing
Either shallow value, so as to obtain the depth angle value of each color in the document, and passes through formula P (c)=Y1A1+Y2A2+Y3A3+Y4A4
+…+YjAjThe user preference value of document color is calculated, wherein:P (c) represents the user preference value of the document color, Y1、
Y2……YjRepresent the depth angle value of each color, A1、A2……AjRepresent that each shade angle value accounts for the area of whole document
Ratio, j >=1;
User preference value P (c) is normalized in the threshold range between the max-thresholds and the minimum threshold, from
And obtain user preference value V2.
7. a kind of user preference analysis method based on document as claimed in claim 6, it is characterised in that by following steps
Calculate the area ratio that each shade angle value accounts for whole document:
To the document, per page carries out sectional drawing, by carrying out the depth that image procossing obtains each pixel of document to every width sectional drawing
Either shallow value;
By the ratio of cumulative for the depth angle value identical pixel rear and total pixel of document, whole text is accounted for as each shade angle value
The area ratio of shelves.
8. a kind of user preference analysis method based on document as claimed in claim 6, it is characterised in that by formula Yj=
Rm+Gb+Bk calculates the depth angle value of each color, wherein:YjRepresent the depth angle value of each color in the document, R, G,
B represents that the numerical value of RGB channel in each pixel, m, b, k represent the coefficient of the RGB channel respectively.
9. a kind of user preference analysis method based on document as claimed in claim 8, it is characterised in that m=0.299, b=
0.587, k=0.114.
10. a kind of user preference analysis method based on document as claimed in claim 3, it is characterised in that by following step
Suddenly extract the document characters in a fancy style information and calculate corresponding user preference value:
The character number of each characters in a fancy style in document is obtained, by formula P (a)=U1C1+U2C2+U3C3…+UtCtCalculate document
The user preference value of characters in a fancy style, wherein:P (a) represents the user preference value of the document characters in a fancy style, U1、U2……UtRepresent described
The character number of each characters in a fancy style, C1、C2……CtThe default preference weight coefficient of expression each characters in a fancy style corresponding, t >=
1;
User preference value P (a) is normalized in the threshold range between the max-thresholds and the minimum threshold, from
And obtain user preference value V3.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610896081.7A CN106503094A (en) | 2016-10-13 | 2016-10-13 | A kind of user preference analysis method based on document |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610896081.7A CN106503094A (en) | 2016-10-13 | 2016-10-13 | A kind of user preference analysis method based on document |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106503094A true CN106503094A (en) | 2017-03-15 |
Family
ID=58294961
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610896081.7A Pending CN106503094A (en) | 2016-10-13 | 2016-10-13 | A kind of user preference analysis method based on document |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106503094A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102426583A (en) * | 2011-10-10 | 2012-04-25 | 北京工业大学 | Chinese medicine tongue manifestation retrieval method based on image content analysis |
CN103106668A (en) * | 2011-11-09 | 2013-05-15 | 佳能株式会社 | Method and system for describing image region based on color histogram |
CN104077344A (en) * | 2013-12-31 | 2014-10-01 | 河南大学 | Self-adaption learning region importance based interactive image retrieval method and system |
-
2016
- 2016-10-13 CN CN201610896081.7A patent/CN106503094A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102426583A (en) * | 2011-10-10 | 2012-04-25 | 北京工业大学 | Chinese medicine tongue manifestation retrieval method based on image content analysis |
CN103106668A (en) * | 2011-11-09 | 2013-05-15 | 佳能株式会社 | Method and system for describing image region based on color histogram |
CN104077344A (en) * | 2013-12-31 | 2014-10-01 | 河南大学 | Self-adaption learning region importance based interactive image retrieval method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110909176B (en) | Data recommendation method and device, computer equipment and storage medium | |
JP7065122B2 (en) | How to get recommended information, devices, electronic devices | |
CN105791157B (en) | A kind of distribution method of flow, distribution system and server | |
CN108021929A (en) | Mobile terminal electric business user based on big data, which draws a portrait, to establish and analysis method and system | |
CN109978630A (en) | A kind of Precision Marketing Method and system for establishing user's portrait based on big data | |
CN108062375A (en) | A kind of processing method, device, terminal and the storage medium of user's portrait | |
CN108984650B (en) | Computer-readable recording medium and computer device | |
CN103955842B (en) | A kind of online advertisement commending system and method towards mass media data | |
CN103793537A (en) | System for recommending individual music based on multi-dimensional time series analysis and achieving method of system | |
CN107222795A (en) | A kind of video abstraction generating method of multiple features fusion | |
CN105426514A (en) | Personalized mobile APP recommendation method | |
CN105142028A (en) | Television program content searching and recommending method oriented to integration of three networks | |
CN103488788A (en) | Method and device for recommending applications automatically | |
CN104899315A (en) | Method and device for pushing user information | |
CN107330020B (en) | User entity analysis method based on structure and attribute similarity | |
CN111177559B (en) | Text travel service recommendation method and device, electronic equipment and storage medium | |
CN104199938B (en) | Agricultural land method for sending information and system based on RSS | |
CN108647818A (en) | A kind of method and device of prediction enterprise concerning taxes risk | |
CN106600213A (en) | Intelligent resume management system and method | |
CN110599393A (en) | Picture style conversion method, device and equipment and computer readable storage medium | |
CN108804577A (en) | A kind of predictor method of information label interest-degree | |
CN108846043A (en) | Network trace mining analysis method and system based on internet big data | |
CN102831219B (en) | A kind of be applied to community discovery can covering clustering method | |
CN106097113A (en) | A kind of social network user sound interest digging method | |
CN104765763B (en) | A kind of semantic matching method of the Heterogeneous Spatial Information classification of service based on concept lattice |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170315 |
|
RJ01 | Rejection of invention patent application after publication |